Please don’t “AI said…”
When someone asks you a question, they’re asking you. Not your chatbot. Not the 47-page document AI just generated for you. You.
If they wanted AI’s take, they’d have asked AI. They probably already did. That’s why they’re asking you.
Don’t do this
— “What do you think about the rollout plan?”
— “Here, I had Claude write this up.” [attaches 12,000 words]
— “Vendor A or Vendor B?”
— “Well, AI said it depends on…”
— “Can you look at this proposal?”
— “I ran it through ChatGPT, here’s the output — thoughts?”
Do this instead
— “What do you think about the rollout plan?”
— “Two phases. Pilot group first, then everyone else. Here’s why: [one paragraph].”
— “Vendor A or Vendor B?”
— “Vendor A. We already work with them — setup will be half the work.”
— “Can you look at this proposal?”
— “Looks good except page 3 — the numbers don’t match the summary on page 1.”
Why this matters
You were asked because you’re the expert in the room. The asker wants your judgment — filtered through your taste, your context, your hard-earned experience.
Forwarding raw AI output does three annoying things at once:
- Makes them read what they could’ve generated in 8 seconds
- Dodges the question — you didn’t take a position, the robot did
- Outsources the thinking they specifically asked you to do
“But AI helped me figure it out!”
Good. Use it. That’s what it’s for.
Then tell me the answer you arrived at. I don’t need the scratchpad, the five rejected options, or the “here are some considerations” bullet list. I need the conclusion, in your voice, that you’re willing to stand behind.
Read it. Disagree with parts. Integrate it with what you actually know. Then come back with a short, specific, committed answer.
That’s what I asked for.